SnapshotReader currently logs tables and rows scanned during the snapshot. It would be useful to expose this data via JMX as well. This will give a better granular view of progress... currently JMX only exposes remainingTableCount, but very large tables can take many hours to complete before this count is decremented.